Barclays Wealth Trustees (Jersey) Ltd & Anor v HM Revenue & Customs

This is an appeal by the taxpayer against a determination of a liability for inheritance tax said to arise in relation to the 10 year charge applicable to trusts. At the heart of the case is the question of whether certain property is “excluded property” for the purposes of the legislation. If it is then the charge could not be made. The demand was made under a notice of determination dated 23rd April 2013. This appeal is made against that determination under section 222(3)(b) of the Inheritance Tax Act 1984 (”the Act” or “the 1984 Act”), pursuant to leave granted by Peter Smith J on 25th November 2014. The facts are short and agreed, and, most unusually in this day and age, the appeal bundle gratifyingly comprised just one physical bundle (including authorities).

The facts are short, but it will help in understanding their significance if I distil the facts and issues to their simplest. Trust property in Trust No 1 was “excluded property”, settled by a non-domiciled settlor, and so would have been free from the 10 year charge had it stayed there. Some of it was transferred to Trust No 2, which had the same settlor who had by now become domiciled in the UK, and it became (at that point) not excluded property. It was then transferred back to Trust No 1. The question, distilled to its simplest, is whether it has re-acquired excluded status.

The actual facts, which were agreed, were as follows. I take this narrative from that set out in the taxpayer’s skeleton argument, with which counsel for HMRC agreed in terms. I have emphasised the key steps by emboldening them.

Thus on the 10th anniversary of the MDT the value of the shares which it had transferred out to the DBJT had come back to it. This appeal is about tax said to be payable in respect of that value. The notice of determination seeks to capture the transferred property (representing the fruits of the Qserv shares transferred back to the MDT) within the 10 year charge applicable to that trust on its 10th anniversary (21st June 2011) and to charge it as non-excluded property. That is where the issue arises. The appellants say that that property was excluded property on that date and therefore not subject to the 10 year charge.

In what follows all references to statute are to the 1984 Act as amended at the relevant dates.

Part III Chapter III of the Act is the chapter in which a periodic charge is imposed on various trusts, including (where otherwise appropriate) the MDT and the DBJT. Section 64, within that Chapter, imposes a 10 year charge to inheritance tax on the MDT where “all or any part of the property comprised in that property is relevant property”. Section 58 describes relevant property for the purposes of this Chapter, and it excludes “… (f) excluded property”. The 10 year period is the 10th anniversary of the “commencement of [the] settlement” - see sections 60 and 61, and under section 60:

“In this Chapter references to the commencement of a settlement are references to the time when property first becomes comprised in it.”

“Excluded property” is dealt with in section 48. This is not part of Chapter III - it is not a definition which is confined to Chapter III, but it applies to it in that “excluded property” is not “relevant property” for the purposes of the 10 year charge. The relevant subsection is subsection (3):

“48 Excluded property

(3)

Where property comprising a settlement is situated outside the United Kingdom –

(a)

the property (but not a reversionary interest in the property) is excluded property unless the settlor was domiciled in the United Kingdom at the time the settlement was made ...”

Thus property will escape the 10 year charge under section 64 if two conditions are fulfilled - the property must be outside the UK, and the settlor must have been domiciled outside the UK when the settlement was made. Those conditions applied to the MDT as originally constituted.

Section 81 governs what happens when property moves from one settlement to another, for the purposes of Chapter III. It has the effect that the property is treated as remaining within the original settlement:

“81 Property moving between settlements

(1)

Where property which ceases to be comprised in one settlement becomes comprised in another then, unless in the meantime any person becomes beneficially entitled to the property (and not merely to an interest in possession in the property), it shall for the purposes of this Chapter be treated as remaining comprised in the first settlement.”

This provision has the effect that the property which moved from the MDT to the DBJT was, at that time, treated as still being within the MDT for the purposes of Chapter III. If that is where it had stayed, the MDT’s 10 year charge period would have applied to it. It was common ground that it would not have retained its character as excluded property in the DBJT because of the effect of section 82 which requires one to look at the domicile of the settlor of the transferee settlement:

“ 82 Excluded property

(1)

For the purposes of this Chapter (except sections 78 and 79) property to which section … 81 above applies shall not be taken to be excluded property by virtue of section 48(3)(a) above unless the condition in subsection (3) below is satisfied (in addition to the conditions in section 48(3) that the property is situated outside the United Kingdom and that the settlor was not domiciled there when the settlement was made).

…

(3)

The condition referred to in subsections (1) or (2) above is – …

b)

in the case of property to which subsection (1) or (2) of section 81 above applies, that the person who is the settlor in relation to the second of the settlements mentioned in the subsection concerned,

was not domiciled in the United Kingdom when that settlement was made.”

Thus for the purposes of the periodic charge (“For the purposes of this Chapter …”) the property is treated as being comprised within the MDT, but in order to see if it has retained its excluded character one has to look to the domicile of the settlor under the second settlement at the time the settlement was made. That settlor was MD, but by this time he had lost his non-domiciled status, so the section 82 condition was not fulfilled. The property lost its excluded character at that point and in that settlement.

Other provisions which are relevant are the definitions of “settlement” and “settlor”:

“43

Settlement and related provisions

(1)

The following provisions of this section apply for determining what is to be taken for the purposes of this Act to be a settlement, and what property is, accordingly, referred to as property comprised in a settlement or as settled property.

(2)

“Settlement” means any disposition or dispositions of property, whether effected by instrument, by parol or by operation of law, or partly in one way and partly in another, whereby the property is for the time being –

(a)

held in trust for persons in succession or for any person subject to a contingency, ….

or would be so held or charged or burdened if the disposition or dispositions were regulated by the law of any part of the United Kingdom; or whereby, under the law of any other country, the administration of the property is for the time being governed by provisions equivalent in effect to those which would apply if the property were so held, charged or burdened.”

“Settlor” is defined in section 44:

“(1)

In this Act ‘settlor’ in relation to a settlement includes any person by whom the settlement was made directly or indirectly, and in particular (but without prejudice to the generality of the preceding words) includes any person who has provided funds directly or indirectly for the purpose of or in connection with the settlement or has made with any other person a reciprocal arrangement for that other person to make the settlement.

(2)

Where more than one person is a settlor in relation to settlement and the circumstances so require, this Part of this Act … shall have effect in relation to it as if the settled property were comprised in separate settlements.”

It is common ground that MD was the settlor in relation to all the movements of the relevant property in this case. It is significant for the purposes of the arguments in this case to note that if there is more than one settlor in relation to a settlement, the respective contributions of the “settlors” are regarded as separate settlements, but there is no express provision treating separate contributions by the same person as being separate settlements.

The effect of the transactions up to the date of the transfer back from the DBJT to the MDT is common ground. It can be summarised as follows:

It is at this point that the cases of the parties diverge.

The case of the taxpayer turns out to be a simple one to express, though elaborating it requires more focus in navigating between the mines of the legislation. It basically says that the returning DBJT property acquires (or reacquires) excluded status because it becomes part of the overall MDT, and at the time that settlement (the MDT) was made MD had not been domiciled in the UK, so section 48(3) makes all the property in the settlement “excluded property”. Elaborating somewhat, section 64 imposes a ten-year anniversary charge where “all or any part of the property comprised in a settlement is relevant property”, and it is charged on the relevant property in the settlement. There is only one settlement in law and in fact for these purposes, and reliance is placed on Rysaffe Trustee Co (CI) Ltd v Inland Revenue Commissioners [2002] STC 872 as establishing that the word “settlement” means what a lawyer would call a settlement. There is no statutory warrant for treating the MDT as two settlements - the original settlement and then a further settlement made of the returning DBJT funds (a proposition said by the taxpayer to lie at the heart of HMRC’s case). The position might have been otherwise if the person introducing those latter funds were a settlor other than Mr Dreelan (the settlor under the MDT) because of the wording of section 44(2), but that was not the case - Mr Dreelan was the settlor in relation to the returning DBJT funds. If one then turns to what the “relevant property” is for the purposes of the 10 year charge, one finds that it does not cover “excluded property”. The effect of section 48(3)(a) is that all the property in “the settlement” is excluded property because “at the time the settlement [viz. the original settlement] was made” Mr Dreelan (the settlor) was not domiciled in the UK. Section 81 does not, of itself, apply specifically to the transfer in from the DBJT because its continuing effect, operating from the occasion of the transfer out to the DBJT, is to deem the funds to have remained in the MDT anyway.

HMRC’s case turns on the event of the transfer back into the MDT. It submits that that event was a disposition effected by an instrument whereby the transferred property was held on the trusts of the MDT. That is said to be an act of settlement for the purposes of determining whether the subject property was excluded property or not. At the date of that act of settlement the settlor (MD) was domiciled in the UK and not outside it. Accordingly the transferred property comprised in that act of settlement was not excluded property (section 48(3)). This analysis involves treating the act of settling the property as an event relevant to the question of “the time the settlement was made” for the purposes of section 48(3) even though the overall settlement into which the money was paid is the entity (settlement) which is taxed, as one settlement, under section 64 (the 10 year charge). In other words, this approach focuses on what I will call the act of settling rather than the resulting legal structure called a settlement.

By way of an alternative argument, HMRC argues that if section 81 has the strong deeming effect relied on by the taxpayers (referred to in the last sentence of paragraph 17 above), then when logically carried through nothing should be treated as having happened on the transfer from the DBJT to the MDT, either in terms of the deemed movement of property or in terms of a change of its status as relevant (not excluded) property. Accordingly, no event has happened to reverse the effect which section 82(3)(b) has already had of removing excluded property status.

Mr Ewart’s argument has a consequence which might be thought to be a little striking, and which would be capable of having an effect beyond the rather arcane facts of the present case. It has the consequence that once a non-domiciled settlor has set up a settlement with some excluded property in it, and if he or she then becomes domiciled in the UK and subsequently adds more foreign property to that settlement, that property has the characteristic of being excluded property (and thus is free from the periodic 10 year anniversary charge) even though the settlor, being by now domiciled in the UK, could no longer set up a new settlement with excluded property. This is because the date for testing domicile is the date the original settlement was made, not when property was subsequently added. This result is even more striking if one imagines a settlement which was seeded with a nominal sum (which frequently happens), with a massive subsequent contribution made when the settlor has become domiciled. Why should that subsequent contribution be able to acquire the characteristic of the original £100 in those circumstances?

Mr Ewart urged me not to be guided by what were submitted to be odd consequences, because an appeal to such arguments was rather less significant once one steps into the world of special charges and deeming which the Act has created. There is much in what Mr Ewart says about that, but the point still has some force. The position is a bit odd. Mr Baldry took me through how the provisions worked in relation to non-domiciled settlors with foreign property, and how it was that those people could settle on trusts without inheritance tax consequences, and how it was that the acquisition or presence of domicile closed off those avenues; and he explained how the stated consequence of Mr Ewart’s argument would be a little at odds with that general background. He did demonstrate a degree of inconsistency, though one probably stopping short of absurdity.

Nonetheless, it is the statute that has to be construed. The key provision in the debate is section 48(3)(a), and the question is what is meant by the phrase “the time the settlement was made”. The taxpayer says that “settlement” means the overall MDT, and the phrase the “time” is the time that that overall settlement was made; the Revenue says “settlement” means the occasion of the property becoming (once again) part of the MDT.

The word “settlement” is important. This provision twice refers to “settlement”. As Mr Ewart pointed out, the word “settlement” in the Act seems generally to be used to describe the concept or structure, and not the act of settling. Although the opening words of section 43(2) (”’Settlement’ means any disposition or dispositions of property …”) might be thought to refer to the act of settling, the rest of the definition (set out above) describes the consequences - the resulting structure, not the act of settling. The Act seems to use the word in that sense extensively - see, for example, sections 43(1), 43(3), 44(2), 47, and 60. Other examples appear. Mr Ewart’s case depends on applying that concept to section 48(3) in a linguistically straightforward way - settlement means the trust structure, and (in this case) the trust structure was created when MD was non-domiciled.

He also submits that that conclusion is consistent with, if not required by, Rysaffe. In that case the question was whether 5 separate settlements fell to be treated as one, as the Revenue contended, invoking the concept of “associated operations”, or as 5, as the taxpayer contended. Park J, with whom the Court of Appeal agreed ([2003] STC 536) considered what “a settlement” was for inheritance tax purposes in paragraphs 18 and following of his judgment. He held that a “settlement” was what a practitioner in the field would think was a settlement, and there were five of those and not just one. That was the position absent any provision in the statute which affected that conclusion, and there was no such provision.

“20.

… However, the Inheritance Tax Act 1984has s 43, which begins by saying that the provisions of the section 'apply for determining what is to be taken for the purposes of this Act to be a settlement'. Notwithstanding those words of the statute, my opinion is that s 43 gives little or no guidance to answering the question of whether, in the many permutations of circumstances of which I gave a few illustrations in [18] above, there is one settlement or more than one. In my judgment the draftsman has for the most part left those questions to be answered in accordance with general principles—in accordance with what I have described above as the general understanding of trust practitioners. It appears that the distinguished editors of Dymond's Capital Taxes are of the same opinion. In para 16.232 they make the point that there are some inheritance tax provisions which enact that in specific circumstances a single settlement under the general law may be more than one settlement for inheritance tax, and (conversely), that there are other inheritance tax provisions which enact that in other specific circumstances a structure which the general law regards as separate settlements may be a single settlement for inheritance tax. But they go on to write: 'In the absence of special provisions, however, the general law must decide [emphasis added].' They clearly do not regard s 43 as a 'special provision' in this sense. In the present case the Revenue do not suggest that any such special provision applies.”

So in ascertaining whether there is a settlement, and what is comprised in it, one looks to the general law. The statute operates on that footing. Mr Ewart’s argument involves applying that concept to section 48(3) and understanding that the “settlement” in paragraph (a) is the MDT, and then concluding, applying ordinary trust principles, that that settlement was made when Mr Dreelan was non-domiciled in the UK, so all the (foreign) property in it, including the property re-acquired from the DBJT, was excluded property on the 10th anniversary.

The contrary argument involves focusing on the word “disposition” in section 43(2). A settlement has to be made by a “disposition”, and Mr Baldry’s case involves identifying the appointment by the DBJT trustees as being a “disposition” which gives rise to a settlement of the property which is subject to it. The property is settled, and it then takes its place in the overall “settlement” (assuming that overall settlement already to exist).

In my view the Revenue’s argument is to be preferred. I reach this conclusion by considering carefully the words in section 48(3)(a) - “the time the settlement was made”. Those words are capable of describing both the making of the original settlement, and the subsequent addition of property to that settlement. The latter is possible because that addition is a disposition whereby the property becomes held on trust, in accordance with the definition in section 43(2). It is not a distortion of the words so to use it. One must not lose sight of the fact that “settlement” is defined in terms which include not merely the end result of something, but which also include the concept of a “disposition”. It is a sort of “disposition plus end result” concept.

It therefore becomes necessary to consider whether that meaning should be preferred, or whether the relevant settlement is the overall settlement into whose funds the new funds are merged. In favour of the latter construction (the taxpayer’s construction) is the fact that the word “settlement” in the opening words of section 48(3)(a) seems, at least at first sight, to refer to the overall settlement, and not the dispositive act of settlement relied on by the Revenue. Mr Baldry conceded that it apparently had that meaning. If that is the meaning of the word in that context then the word would have two different meanings in the same subsection, and on normal principles of statutory interpretation that is unlikely. For my part I do not think that that first reference necessarily does have that meaning. The focus of the subsection is on “property”, and if the “property” is to be comprised in a “settlement” it must have got there as a result of a “disposition”. A good case can therefore be made for saying that the first reference is a reference to the “disposition plus trust” definition of settlement, rather than the end structure.

However, in the light of the concession made by Mr Baldry I shall not pursue that line. I shall assume that the first reference is simply a reference to the existing structure (here, the MDT), and not to the act of settling the property by reference to which the inquiry is made. I still nonetheless consider that the reference to the “making” of the “settlement” is a reference to the occasion on which the relevant property became settled, notwithstanding that that deploys a different meaning to “settlement” than the opening words of the subsection. The “settlement” which is made is the “disposition plus trust structure” which affected the property when it was appointed by the trustees of the DBJT. The verb “made” is, in my view, significant. When the draftsman intended to refer to the initial creation of the overall settlement he used different wording in sections 60 and 61 (“commencement”and “the time when property first becomes comprised in it”). If one is looking to that occasion the word “made” is not so appropriate. The word “made” suggests an act, and not just an end result.

This idea occurs elsewhere. In section 44(1) the word “made” is used to connote something which focuses more on the act of disposition of property. The words “ … any person by whom the settlement was made” and its following words suggest a reference to dispositions (and therefore a settlement in the “disposition plus end result” sense). Similarly in section 48(3)(a) the word “made”, coupled with the fact that “property” is a focus, is a strong pointer towards the “disposition plus trust structure” meaning of the second reference to “settlement”. In effect, the words should be read as if they said “at the time the settlement of the property was made”. Thus when the property was returned to the MDT, MD was the “settlor” because he “made” the settlement of it within the meaning of section 44(1). There was a “disposition … whereby the property was held on trust” within the meaning of section 43. It is therefore entirely consistent to say that for the purposes of section 48(3)(a) the settlement of that property was “made” when it accrued to the MDT.

This view is supported by logic and plausibility in overall taxation terms. I can detect no logic, even in an artificially constructed tax world, for allowing excluded property status to after-contributed property by a domiciled settlor merely because when he created the settlement in the first place he was non-domiciled. I do not consider it plausible that Parliament had that intention.

It should be noted that the structure of the Act would not allow that consequence where the subsequent contributor was a domiciled third party (another settlor). In those circumstances section 44(2) provides that contributions by such a person should be treated as a separate settlement, so if that person is domiciled in the UK, his or her contribution plainly cannot be excluded property on any construction of section 48(3). Mr Ewart acknowledged this contrast, but seemed to suggest that the contrast between this express effect and the fact that a same-settlor contribution is not expressed to give rise to a separate settlement supported his construction. I draw the contrary conclusion. The fact that a contribution from a domiciled third party settlor would not be excluded property demonstrates the anomaly of allowing that status to additions from the original settlor when domiciled. That conclusion would not mean there is no distinction between contributions by third parties and additional contributions by the settlor. The former would still be treated as a separate settlement for the purposes of, for example, the 10 year charge, whereas the latter would not.

Part of Mr Ewart’s answer to the Revenue’s reliance on a disposition on the occasion of the passage of the property from the DBJT to the MDT lies in section 81. That section had a deeming effect when the appointment from the MDT to the DBJT took place. Under that section the property was treated as “remaining comprised in the first settlement [ie the MDT]”. His submission was that since it was treated as being there already at the time of the actual re-transfer from the DBJT, there could be no “disposition” for the purposes of the relevant provisions of the Act. Accordingly the Revenue did not have the benefit of the “disposition” that underpinned its argument.

In support of his argument Mr Ewart cited Peter Gibson J in Marshall (Inspector of Taxes) v Kerr [1993] STC 360, cited with apparent approval in DV3 RS Ltd Partnership v Revenue and Customs Commissioners [2013] STC 2150 at para 13:

“For my part I take the correct approach in construing a deeming provision to be to give the words used their ordinary and natural meaning, consistent so far as possible with the policy of the Act and the purposes of the provisions so far as such policy and purposes can be ascertained; but if such construction would lead to injustice or absurdity, the application of the statutory fiction should be limited to the extent needed to avoid such injustice or absurdity, unless such application would clearly be within the purposes of the fiction. I further bear in mind that because one must treat as real that which is only deemed to be so, one must treat as real the consequences and incidents inevitably flowing from or accompanying that deemed state of affairs, unless prohibited from doing so.”

It is the last sentence of that citation that is relied on by Mr Ewart. He starts with the (deemed) fact that the property appointed from the MDT to the DBJT, and then back again, was, by virtue of section 81, deemed to have remained in the MDT, and says that it is an inevitable consequence, or incident, of that that there was no disposition on the second of those transfers. I do not accept that submission. The deeming provision in section 81 has to be considered carefully. It operates for the purposes of Chapter III, and only for the purposes of that Chapter. It has important effects in relation to how the periodic charge works. However, it would be going far too far to say that one inevitably has to treat a real world disposition as though it had not taken place at all. That is not necessary or inevitable. The taxation consequences of the deeming, so far as they are within Chapter III, might be inevitable, but the absence of a real world disposition is not.

One can usefully look at the position as at the date of the MDT/DBJT transfer. On that occasion there was a disposition. Unless there had been section 81 would have no work to do. Section 81 provides some deeming effects despite that disposition. It is not necessary to assume that the disposition never happened. One just ignores its dispositive effect for the purposes of Chapter III. Accordingly one does not deem there to have been no disposition on that occasion. It is no more appropriate to deem there to be no disposition on the second (DBJT/MDT) transfer. The funds were still treated, for the purposes of Chapter III, as being in the MDT, but it is not necessary to go further and deem there to have been no disposition.

Furthermore, Mr Ewart’s submission would result in section 81 operating outside the scope of Chapter III. Section 48 is not in Chapter III. The concept of excluded property is capable of operating across the whole statute. It is deployed within Chapter III (because section 58 prevents “excluded property” from being “relevant property” for the purposes of the Chapter) but that does not bring it within Chapter III. Accordingly, if Mr Ewart were correct he would be giving section 81 a wider reach than that which was intended for it - it would be reaching into Chapter II.

Mr Ewart also pointed to what he said was a further inconsistency in the position of the Revenue. He sought to say that if the Revenue was right in saying that the passing of property from the DBJT to the MDT was a settlement for the purposes of section 48(3) then the Revenue ought to acknowledge the consequence of that, which was that the property was not part of the original settlement for the purposes of the periodic charge relied on by the Revenue. The Revenue’s demand, based on the periodic charge, assumed that the commencement date of the relevant settlement was the date of the commencement of the MDT. But if the property which rejoined from the DBJT was the subject of a separate settlement, then it did not fall to be treated as part of the MDT at that time. According to Mr Ewart it ought, logically, to be taxed under Chapter III as if it were a separate settlement. However, the Revenue’s reasoning did not provide for that, and was faulty for that reason. Mr Ewart actually advanced this argument as a reason why the Revenue could not be correct in its submissions about excluded property, and not (or not clearly) as a standalone objection to the notice of determination, but in the light of my finding about the correctness of the Revenue’s submissions it becomes a point that I have to deal with, as it seems to me.

I do not consider that Mr Ewart’s argument is correct. The true construction of section 48(3) is one that requires one to look at the occasion of the settling of the property for the purposes of determining whether or not it is excluded property, and nothing else. It does not create a separate settlement for the other purposes of the Act, deemed or otherwise. The overall settlement for the purposes of section 64 remains the same. Rysaffe enables (and requires) one to look at basic trust law for determining that. Section 64 imposes a charge on the relevant property “comprised in a settlement” on the ten year anniversary; section 61 measures that 10 year anniversary from the “date on which the settlement commenced”; and section 60 provides that:

“References to the commencement of a settlement are to the time when property first became comprised in it.”

All those are (in the circumstances) references to the MDT, implicitly acknowledging that property may have arrived in it at different times. That is what has happened here. The correct interpretation of section 48(3) has no impact on this. It involves considering the circumstances of the settlement of property at one point in time, and does not otherwise involve the creation of any “settlement” which differs from the overall MDT.

Nor does Rysaffe itself compel a contrary conclusion on section 48(3). That case determines that the concept of what is a settlement, for the purposes of the provisions of the Act considered by Park J, is the same as that understood by the general law. Thus in his case there were 5 real world settlements, and not one, and therefore there were 5 settlements for the purposes of the Act. Park J was not considering every reference to the word “settlement” in the Act, and was not considering the detail of what was the making of a settlement for the purposes of section 48(3)(a). I therefore consider that Rysaffe does not assist Mr Ewart.

The conclusion which I have reached in relation to the operation of section 48(3) makes it unnecessary for me to consider the Revenue’s alternative point. It is a point which it is not easy to get hold of as a point, and dealing with it involves the unrewarding and unremitting pursuit of deeming provisions to a degree which might be thought to be questionable. Since it is not necessary for me to consider it, since it is far from straightforward, and since anything I say may affect other situations which are not before the court, I shall not express any views on it.

It follows that this appeal stands to be dismissed.